Converting patient-reported physical function outcomes scores to promis metric scores in phase 3 trials of baricitinib in rheumatoid arthritis

Background: In patients (pts) with RA, physical function (PF) can be measured with the Health Assessment Questionnaire-Disability Index (HAQ-DI). Patient- Reported Outcomes Measurement Information System (PROMIS) was developed by the National Institutes of Health using a population-calibrated T-score metric (Mean 50, SD 10). Crosswalk tables that link legacy instruments to PROMIS instruments, including PF, have been developed. Comparisons to the general population can also be made from PROMIS scores.1-2 Objectives: To convert HAQ-DI scores to PROMIS PF scores to determine how the PROMIS metric performs in 2 phase 3 baricitinib (bari) clinical trials in pts with RA. Methods: In RA-BEAM, pts with inadequate response (IR) to methotrexate were randomised 3:3:2 to placebo (PBO) once daily (QD), bari 4 mg QD, or adalimumab (ADA) 40 mg biweekly.3 In RA-BEACON, pts with IR to bDMARDs were randomised 1:1:1 to receive PBO or bari 2 mg or 4 mg QD.4 In both studies, PF was assessed using HAQ-DI. Patient-level HAQ-DI scores were converted to PROMIS PF scores using a validated crosswalk table.1 Analysis of covariance was conducted on the PROMIS PF score conversions to compare bari to all treatment arms in both studies. Missing data were imputed using modified last observation carried forward. Results: Pts had considerable PF impairment at baseline; mean scores exceeded 2 SD (20 points on the T-score metric) from population means (table 1). Treatment with bari was associated with clinically relevant improvements approaching or exceeding 0.5 SD (5 points on the T-score metric) by week 24 (minimally important difference for PROMIS PF.: 0.2 SD or 2 points5) vs PBO in HAQ-DI converted to PROMIS PF scores (table 1). Using the converted PROMIS scores, bari remained associated with significant improvements in PF vs PBO through 24 weeks in both studies and vs ADA through 52 weeks in RA-BEAM (figure 1). Conclusions: While RA-BEAM and RA-BEACON did not use the PROMIS instrument directly, these results indicate PROMIS PF has the potential to show responsiveness and differentiate between active treatments in clinical trials.